Using Tumor Microenvironment Mapping to Predict Lung Cancer Immunotherapy Outcomes

In a recent groundbreaking study, researchers unveiled how analyzing the spatial organization of immune cells within lung tumors can improve predictions of patient responses to immunotherapy. This research highlights that the way immune cells cluster around tumor cells—known as tumor microenvironments—can serve as a more reliable biomarker than traditional tests like PD-L1 expression or tumor mutational burden.

Lung cancer remains the leading cause of cancer-related deaths worldwide, with non-small-cell lung cancer (NSCLC) constituting over 80% of cases. Despite advances with immune checkpoint inhibitors, only about 27–45% of patients experience meaningful benefits. Current predictive biomarkers are often inconsistent, prompting the need for more precise methods.

The study, published in Science Advances, involved an integrated approach combining multiplex immunofluorescence (mIF), RNA sequencing, and deep-learning–based histology. Researchers examined tissue samples from 132 NSCLC patients treated at Stanford Medical Center—50 with detailed mIF imaging, 115 with histology slides, and 122 with RNA data—profiling over 45 million individual cells.

Using sophisticated clustering and deep-learning models, scientists identified distinct tumor-immune architecture patterns. They developed a cytotoxic T lymphocyte (CTL) score reflecting the density of cytotoxic T cells in tumor neighborhoods. Responders to PD-1/PD-L1 blockade therapy exhibited significantly higher CTL scores, with 2.5 times more cytotoxic T cells and 6.5 times more Tc-enriched neighborhoods compared to nonresponders. These tumors also showed stronger interactions between immune cells and tumor cells.

Furthermore, patients with higher CTL scores experienced longer progression-free survival. Notably, tumors in former smokers exhibited fewer immune cell neighborhoods directly neighboring cancer cells, whereas never smokers had denser lymphocyte clusters adjacent to tumor tissues.

This research suggests that integrating detailed spatial immune profiling into routine pathology could greatly enhance patient selection for immunotherapy, leading to better outcomes and reduced unnecessary toxicity. The potential to apply these insights using software like NucSegAI, which can analyze standard H&E slides, offers a promising pathway toward more personalized lung cancer treatment.

Source: https://medicalxpress.com/news/2025-05-tumor-microenvironments-lung-cancer-immunotherapy.html